A data fusion method for a distributed energy supply system

Abstract

The present invention relates to the technical field of power line interference removal, and specifically relates to a data fusion method for a distributed power supply system, including the following steps: Step 1, regularly poll designated registers of each protocol device through the data fusion terminal network port; Step 2, use a highly selective elliptic function adaptive notch filter set in the processor to suppress power line interference from short-latency evoked potentials through time-varying stopband width and non-zero initial values, and filter all input sampled data; Step 3, through the adaptive Ethernet interface of the processor, fuse the filtered data into corresponding messages according to preset requirements and upload them to the background system. This method can simultaneously take into account the optimization problems of the transient response duration and amplitude.

Description

Technical Field

[0001] The present invention relates to the technical field of power line interference removal, and particularly to a data fusion method for a distributed energy supply system. Background Art

[0002] Power line (electric field line) interference is one of the common interferences affecting data fusion terminals in distributed energy supply systems, and its effective removal remains a challenge in distributed energy supply systems. Power line interference is usually characterized by a fixed-frequency sine wave (50 / 60 Hz, depending on the country / region) with random phase and amplitude and higher harmonics.

[0003] Existing studies have proposed many signal processing methods and algorithms to solve the problem of power line interference removal. Among them, notch filtering is the most commonly used method for removing power line noise, and a series of multi-notch filters (fixed notch filters) are centered on the fundamental frequency and higher harmonics of the power line. The multi-notch filter can be implemented as an FIR or IIR structure. Compared with an FIR filter that satisfies the same frequency assumption, the order of the IIR filter can be designed to be much smaller. Among them, the cascading method is a direct design method for IIR multi-notch filters, and the multi-notch filter is designed by cascading the required number of second-order single-notch filters; while other methods are usually based on the optimization of pole positions and require the use of advanced iterative algorithms or complex search techniques. The multi-notch filter can suppress the selected frequencies in the input signal. If the notch (selected frequency) is narrow enough, the attenuation of other frequencies can be minimized; the bandwidth of the notch directly depends on the pole radius of the filter. As the pole radius approaches the unit circle, it becomes more selective. However, narrowing the notch of the filter will increase the duration of the transient response of the filter. Therefore, on the one hand, the interference suppression filter should be very selective and have a narrow band, and at the same time it should have good time-domain characteristics. Choosing between a very sharp amplitude response, a narrow notch, and a short transient response duration is one of the important issues in designing the optimal multi-notch filter.

[0004] In summary, how to simultaneously consider the optimization problems of the transient response duration and amplitude has become an urgent problem to be solved for power lines in distributed energy supply systems for data fusion. Summary of the Invention

[0005] Aiming at the deficiencies of the above-mentioned existing technologies, the present invention provides a data fusion method for a distributed energy supply system, which can simultaneously consider the optimization problems of the transient response duration and amplitude.

[0006] To solve the above technical problems, the present invention adopts the following technical solutions:

[0007] A data fusion method for a distributed energy supply system includes the following steps:

[0008] Step 1: Regularly poll the specified registers of each protocol device through the data fusion terminal network port;

[0009] Step 2: Use the high-selectivity elliptic function adaptive notch filter set in the processor to suppress the power line interference from short-latency evoked potentials through time-varying stopband width and non-zero initial value, and filter all the input sampled data;

[0010] Step 3: Through the adaptive Ethernet interface of the processor, after fusing the filtered data into corresponding messages according to preset requirements, upload them to the background system;

[0011] Among them, in Step 2, the design process of the high-selectivity elliptic function adaptive notch filter includes:

[0012] S1: Synthesize multiple prototype notch filters into a prototype multi-notch filter; the multiple prototype notch filters are all elliptic filters;

[0013] S2: Introduce a non-zero initial value to the prototype multi-notch filter to suppress the transient response of the prototype multi-notch filter;

[0014] S3: Introduce a time-varying stopband width into the prototype multi-notch filter to obtain a high-selectivity elliptic function adaptive notch filter.

[0015] Preferably, in S2, the calculation process of the non-zero initial value includes:

[0016] S201: Calculate the non-zero initial value vector

[0017]

[0018] where I is the identity matrix, P is the projection operator, and X is the vector of input samples;

[0019] S202: Analyze the correlation between the filtering quality measured by the mean square error MSE and the length k of the initial value vector to obtain the k value that makes the initial value of the selected filter have the best interference suppression effect;

[0020] S203: Obtain the non-zero initial value according to the optimal k value and the initial value vector

[0021] Preferably, in S201, the calculation formula of the projection operator P is:

[0022] P = B(B T B) -1 B T ;

[0023] ​

[0024] Among them, ω represents the power supply interference frequency, and h is the number of the highest harmonics.

[0025] Preferably, the calculation formula of the vector X of the input sample is:

[0026] X = U + D = [x(0) x(1) … x(k - 1)] T ;

[0027]

[0028] x(n) = u(n) + d(n);

[0029]

[0030] Among them, x(n) is the input signal, u(n) is the ideal signal, d(n) is the sinusoidal interference of the known frequency, i is the exponent of the harmonic, h is the number of the highest harmonics, A i is the amplitude of the harmonic, ω is the power supply interference frequency, φ i is the phase of the harmonic.

[0031] Preferably, in S202, the calculation formula of the mean square error MSE is:

[0032]

[0033] In the formula, N is the window size, and y(n) is the filtered output signal.

[0034] Preferably, in S3, after introducing the time-varying stopband width, the description formula of the high-selectivity elliptic function adaptive notch filter is:

[0035]

[0036] In the formula, y(n) represents the filtered signal, l is the length of the change range in the sample, b 0m (n), …, b Nm (n) and a 1m (n), …, a Nm (n) are the coefficients in the time-domain equation corresponding to the transfer function of the prototype multi-notch filter.

[0037] Preferably, in S1, multiple prototype notch filters are all centered on the power line fundamental frequency of 60 Hz or higher harmonics.

[0038] Preferably, in S1, the number of prototype notch filters is five; the center frequencies of the five prototype notch filters are respectively: ω 1 = 60 Hz, ω 2 = 180 Hz, ω 3 = 540 Hz, ω4 = 900 Hz, ω 5 = 1260 Hz.

[0039] Preferably, the maximum passband ripple, minimum stopband attenuation, and stopband width of the five prototype notch filters are the same, and the maximum passband ripple Rp = 0.1 dB, the minimum stopband attenuation Rs = 40 dB, and the stopband width BW = 4 Hz.

[0040] Preferably, the transfer function of the prototype multi-notch filter is:

[0041]

[0042]

[0043] where H 1 (z), H 2 (z), H 3 (z), H 4 (z), H 5 (z) are the transfer functions of the five prototype notch filters respectively, z -1 , z -2 , …, z -30 are the unit impulse response sequences of the prototype notch filters, a 1 , a 2 , …, a 30 and b 0 , b 1 , b 2 , …, b 30 are the coefficients corresponding to z 1 (z), H 2 (z), H 3 (z), H 4 (z), H 5 (z) after multiplication and calculation for the corresponding z -1 , z -2 , …, z -30 before.

[0044] Compared with the prior art, the present invention has the following beneficial effects:

[0045] 1. In the present invention, the data of the protocol device collected will be filtered by the high-selectivity elliptic function adaptive notch filter in the processor and then the message will be uploaded.

[0046] The high - selectivity elliptic - function adaptive notch filter consists of multiple elliptic filters. The elliptic filter provides a sharp transition band, thus effectively eliminating the narrow - band power - line noise in the evoked potential and separating the noise from the useful signal. The elliptic filter has a high quality factor, a sharp transition band, and sufficient selectivity, and can effectively eliminate the narrow - band power - line noise in the evoked potential. Compared with Butterworth filters or Chebyshev filters, the elliptic filter can provide better suppression. The high - selectivity elliptic - function adaptive notch filter composed of multiple elliptic filters, that is, a digital IIR multi - notch filter based on an elliptic - prototype analog filter, has a transient - response duration and amplitude that are significantly lower than those of traditional filters and can more effectively eliminate power - line interference. In addition, during the design process of the high - selectivity elliptic - function adaptive notch filter, a non - zero initial value is introduced. Compared with filters using zero initial values, it can further eliminate power - line interference. And a time - variable stop - band width is introduced into the structure corresponding to the second stage of the transient - suppression process. Since the transient - response duration of the filter depends on the notch bandwidth, narrowing the notch of the filter will increase the transient - response duration of the filter. After introducing the time - variable stop - band width, the transient - response duration of the filter can be effectively reduced.

[0047] In summary, the transient - response duration and amplitude of the present invention are significantly lower than those of traditional filters, and it can take into account the optimization problems of both the transient - response duration and amplitude.

[0048] 2. In the present invention, a design process of a high - selectivity elliptic - function adaptive notch filter is provided. Through the calculation process of the non - zero initial value provided by the present invention, the corresponding non - zero initial value can be accurately, quickly, and marginally obtained.

[0049] 3. The present invention also provides a preferred method for selecting a prototype notch filter, including the number of selected prototype notch filters, center frequency, maximum pass - band ripple, minimum stop - band attenuation, and stop - band width. By referring to the prototype notch filter provided by the present invention and combining the processing methods of the non - zero initial value and the time - variable stop - band width in the present invention, a high - selectivity elliptic - function adaptive notch filter with a transient - response duration and amplitude significantly lower than those of traditional filters can be obtained. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] In order to make the objectives, technical solutions, and advantages of the invention clearer, the present invention will be further described in detail below with reference to the drawings, where:

[0051] Figure 1 is the structural block diagram of the data - fusion terminal for the distributed energy - supply system in the embodiment;

[0052] Figure 2 is the flow chart of the data - fusion method for the distributed energy - supply system in the embodiment;

[0053] Figure 3 Flow chart of the design process of the high - selectivity elliptic - function adaptive notch filter in the embodiment;

[0054] Figure 4 Schematic diagram of the amplitude and time responses of the prototype multi - notch filter in the embodiment;

[0055] Figure 5 Schematic diagram of the values of some coefficients in the embodiment;

[0056] Figure 6 For the mean square error and the initial value vector Schematic diagram of the correlation between the length k in the embodiment;

[0057] Figure 7 Schematic diagram of the time responses of the 4 - Hz and 30 - Hz bandwidths of the prototype multi - notch filter to a 60 - Hz sinusoidal input signal in the embodiment. Detailed implementation manners

[0058] The following is a more detailed description through specific implementation manners:

[0059] Embodiment:

[0060] It should be noted that for the convenience of understanding, the block diagram of the data fusion terminal for the distributed energy supply system in this embodiment is as Figure 1 shown. The data fusion terminal for the distributed energy supply system includes a Cortex - a7 processor, 4 groups of RS485 ports, 20 MODBUS - protocol devices, a 10 / 100M adaptive Ethernet interface, Ethernet, and a power supply.

[0061] The specific application scenario in this embodiment is as follows: The background system periodically polls the specified registers of 20 MODBUS - protocol devices connected to 4 groups of RS485 ports through the data fusion terminal network interface; then, the Cortex - a7 processor filters all the input sampled data; finally, all the collected data is fused into a specific message according to requirements through the 10 / 100M adaptive Ethernet interface of the processor and actively uploaded to the background system at regular intervals. Specifically, when implementing, the main frequency of the Cortex - a7 processor is 1.2 GHz, each RS485 has an independent thread, and the network has an independent thread; each group of RS485 interfaces independently sets communication parameters, the baud rate supports 1200 - 115200 bp / s, and the shortest polling interval time is 1 minute; 4 groups of RS485 buses, independently configured to automatically collect registers, and each register is separately configured with a data conversion method (fixed - point to floating - point, multi - byte arrangement order); the processor supports network interface upgrade.

[0062] As Figure 2As shown in the figure, a data fusion method for a distributed energy supply system is disclosed in this embodiment, including the following steps:

[0063] Step 1: Regularly detect the specified registers of each protocol device through the data fusion terminal network port;

[0064] Step 2: Use the high-selectivity elliptic function adaptive notch filter set in the processor to suppress the power line interference from short-latency evoked potentials through time-varying stopband width and non-zero initial value, and filter all the input sampled data;

[0065] Step 3: Through the adaptive Ethernet interface of the processor, after fusing the filtered data into corresponding messages according to preset requirements, upload them to the background system;

[0066] Among them, as Figure 3 shown in the figure, in Step 2, the design process of the high-selectivity elliptic function adaptive notch filter includes:

[0067] S1: Synthesize multiple prototype notch filters into a prototype multi-notch filter; the multiple prototype notch filters are all elliptic filters.

[0068] As Figure 4 shown in the figure, in this embodiment, the number of prototype notch filters is five; the center frequencies of the five prototype notch filters are respectively: ω 1 = 60Hz, ω 2 = 180Hz, ω 3 = 540Hz, ω 4 = 900Hz, ω 5 = 1260Hz; the maximum passband ripple, minimum stopband attenuation and stopband width of the five prototype notch filters are the same, and the maximum passband ripple Rp = 0.1dB, the minimum stopband attenuation Rs = 40dB, and the stopband width BW = 4Hz.

[0069] The transfer function of the prototype multi-notch filter is:

[0070]

[0071]

[0072] Among them, H 1 (z), H 2 (z), H 3 (z), H 4 (z), H 5 (z) are the transfer functions of the five prototype notch filters respectively, z -1 , z -2 , …, z -30 is the unit impulse response sequence of the prototype notch filter, a1 、 a 2 、…、 a 30 and b 0 、 b 1 、 b 2 、…、 b 30 are H 1 (z), H 2 (z), H 3 (z), H 4 (z), H 5 (z) are multiplied and then the coefficients corresponding to z -1 、 z -2 、…、 z -30 before are calculated.

[0073] In this embodiment, the values of a1, a2, …, a30 and b0, b1, b2, …, b30 are as Figure 5 shown.

[0074] S2. Introduce non - zero initial values to the prototype multi - notch filter to suppress the transient response of the prototype multi - notch filter.

[0075] The calculation process of the non - zero initial values includes:

[0076] S201. Calculate the non - zero initial value vector

[0077]

[0078] where I is the identity matrix, P is the projection operator, and X is the vector of input samples;

[0079] The calculation formula of the projection operator P is:

[0080] P = B(B T B) -1 B T ;

[0081]

[0082] where ω represents the power interference frequency and h is the number of the highest harmonics.

[0083] The calculation formula of the vector X of input samples is:

[0084] X = U + D = [x(0) x(1) … x(k - 1)] T ;

[0085]

[0086] x(n) = u(n) + d(n);

[0087]

[0088] where \(x(n)\) is the input signal, \(u(n)\) is the ideal signal, \(d(n)\) is a sinusoidal interference with a known frequency, \(i\) is the exponent of the harmonic, \(h\) is the number of the highest harmonics, \(A\) i is the amplitude of the harmonic, \(\omega\) is the power supply interference frequency, \(\varphi\) i is the phase of the harmonic.

[0089] S202. Analyze the correlation between the filtering quality measured by the mean square error MSE and the length \(k\) of the initial value vector to obtain the \(k\) value that makes the interference suppression effect of the selected filter initial value optimal;

[0090] where the calculation formula of the mean square error MSE is:

[0091]

[0092] In the formula, \(N\) is the window size, and \(y(n)\) is the filtered output signal.

[0093] In this embodiment, as Figure 6 shown, when \(k\) is an integer multiple of \(f\) s / \(f\) 0 , MSE reaches a local minimum; In summary, the local minimum \(k\) of the function MSE(\(k\)) min can be determined by the following relationship:

[0094]

[0095] In the formula, \(f\) s is the sampling rate, \(f\) 0 is the power supply interference frequency, \(j\) is an integer; The selection of the length \(k\) of the initial condition vector is a trade-off between the filtering quality and the delay caused by the need to record \(k\) samples in the signal \(x(n)\).

[0096] S203. Obtain non-zero initial values according to the optimal \(k\) value and the initial value vector .

[0097] S3. Introduce a time-varying stopband width into the prototype multi-notch filter to obtain a high-selectivity elliptic function adaptive notch filter.

[0098] As Figure 7 shown, the transient response duration of the filter depends on the bandwidth of the notch. Narrowing the notch will increase the transient response duration of the filter. Therefore, in the second stage of the transient suppression process of the multi-notch filter under consideration, a time-varying stopband width is introduced into the filter structure.

[0099] After introducing the time-varying stopband width, the description formula of the high-selectivity elliptic function adaptive notch filter is:

[0100]

[0101] Wherein, y(n) represents the filtered signal, l is the length of the variation range in the sample, and b 0m (n), …, b Nm (n) and a 1m (n), …, a Nm (n) are the coefficients in the time-domain equation corresponding to the transfer function of the prototype multi-notch filter.

[0102] The linear function of the stopband width length of the filter is the simplest form of the change in the stopband width; in this embodiment, to ensure that the filter operation has sufficient speed at the beginning, the initial value BW of the stopband width start = 30 Hz; the final stopband width BW is determined by the established filter selectivity final = 4 Hz.

[0103] In the present invention, the data of the protocol device collected will be filtered by the high-selectivity elliptic function adaptive notch filter in the processor and then the message will be uploaded. The high-selectivity elliptic function adaptive notch filter is composed of multiple elliptic filters. The elliptic filter provides a sharp transition band, thus effectively eliminating the narrowband power line noise in the evoked potential and separating the noise from the useful signal; the elliptic filter has a high quality factor, a sharp transition band and sufficient selectivity, and can effectively eliminate the narrowband power line noise in the evoked potential; compared with the Butterworth filter or the Chebyshev filter, the elliptic filter can provide better suppression. The high-selectivity elliptic function adaptive notch filter composed of multiple elliptic filters, that is, the digital IIR multi-notch filter based on the elliptic prototype analog filter, has a transient response duration and amplitude that are significantly lower than those of the traditional filter, and can more effectively eliminate the power line interference. In addition, during the design process of the high-selectivity elliptic function adaptive notch filter, a non-zero initial value is introduced. Compared with the filter using a zero initial value, it can further eliminate the power line interference; and a time-varying stopband width is introduced in the structure corresponding to the second stage of the transient suppression process. Since the transient response duration of the filter depends on the notch bandwidth of the filter, narrowing the notch of the filter will increase the transient response duration of the filter. After introducing the time-varying stopband width, the transient response duration of the filter can be effectively reduced.

[0104] In addition, in the present invention, a design process of a high-selectivity elliptic function adaptive notch filter is provided. Through the calculation process of the non-zero initial value provided by the present invention, the corresponding non-zero initial value can be accurately, quickly and marginally obtained. The present invention also provides a preferable selection method for a prototype notch filter, including the number, center frequency, maximum passband ripple, minimum stopband attenuation and stopband width of the selected prototype notch filter. By referring to the prototype notch filter provided by the present invention and combining the processing method of the non-zero initial value and the time-varying stopband width in the present invention, a high-selectivity elliptic function adaptive notch filter with a transient response duration and amplitude significantly lower than those of traditional filters can be obtained.

[0105] In summary, the transient response duration and amplitude of the present invention are significantly lower than those of traditional filters, and the optimization problems of both the transient response duration and amplitude can be taken into account simultaneously.

[0106] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit them. Those of ordinary skill in the art should understand that any modifications or equivalent replacements made to the technical solutions of the present invention without departing from the spirit and scope of the present technical solution shall be covered by the scope of the claims of the present invention.

Claims

1. A data fusion method for a distributed energy supply system, characterized in that, it includes the following steps: Step 1: Regularly detect the specified registers of each protocol device through the data fusion terminal network port; Step 2: Use the high-selectivity elliptic function adaptive notch filter set in the processor to suppress the power line interference from short-latency evoked potentials through time-varying stopband width and non-zero initial value, and filter all input sampled data; Step 3: Through the adaptive Ethernet interface of the processor, after fusing the filtered data into corresponding messages according to preset requirements, upload them to the background system; Among them, in Step 2, the design process of the high-selectivity elliptic function adaptive notch filter includes: S1: Synthesize multiple prototype notch filters into a prototype multi-notch filter; the multiple prototype notch filters are all elliptic filters; S2: Introduce a non-zero initial value to the prototype multi-notch filter to suppress the transient response of the prototype multi-notch filter; S3: Introduce a time-varying stopband width into the prototype multi-notch filter to obtain a high-selectivity elliptic function adaptive notch filter; Among them, in S2, the calculation process of the non-zero initial value includes: S201. Calculate a non-zero initial value vector where I is the identity matrix, P is the projection operator, and X is the vector of input samples; S202. Analyze the correlation between the filtering quality measured by the mean square error MSE and the length k of the initial value vector to obtain the k value that optimizes the interference suppression effect of the selected filter initial value; S203. Obtain a non-zero initial value according to the optimal k value and the initial value vector Obtain a non-zero initial value; In S3, after introducing the time-varying stopband width, the description formula of the high-selectivity elliptic function adaptive notch filter is: Where, x(n) is the input signal, y(n) represents the filtered signal, l is the length of the variation range in the samples, and b 0m (n), …, b Nm (n) and a 1m (n), …, a Nm (n) are the coefficients in the time-domain equation corresponding to the transfer function of the prototype multi-notch filter.

2. The data fusion method for a distributed energy supply system according to claim 1, characterized in that: In S201, the calculation formula of the projection operator P is: P = B(B T B) -1 B T ; where ω represents the power supply interference frequency and h is the number of the highest harmonics.

3. The data fusion method for a distributed energy supply system according to claim 2, characterized in that: The calculation formula of the vector X of input samples is: X = U + D = [x(0) x(1) … x(k - 1)] T ; U = [u(0) u(1) … u(k - 1)] T D = [d(0) d(1) … d(k-1)] T ; x(n) = u(n) + d(n); Among them, x(n) is the input signal, u(n) is the ideal signal, d(n) is the sinusoidal interference with a known frequency, i is the exponent of the harmonic, h is the number of the highest harmonics, A i is the amplitude of the harmonic, ω is the power supply interference frequency, φ i is the phase of the harmonic.

4. The data fusion method for a distributed energy supply system according to claim 3, characterized in that: In S202, the calculation formula of the mean square error MSE is: In the formula, N is the window size and y(n) is the filtered output signal.

5. The data fusion method for a distributed energy supply system according to claim 4, characterized in that: In S1, multiple prototype notch filters are all centered on the power line fundamental frequency of 60Hz or higher harmonics.

6. The data fusion method for a distributed energy supply system according to claim 5, characterized in that: In S1, the number of prototype notch filters is five; the center frequencies of the five prototype notch filters are respectively: ω 1 = 60 Hz, ω 2 = 180 Hz, ω 3 = 540 Hz, ω 4 = 900 Hz, ω 5 = 1260 Hz.

7. The data fusion method for a distributed energy supply system according to claim 6, characterized in that: The maximum passband ripple, minimum stopband attenuation and stopband width of the five prototype notch filters are the same, and the maximum passband ripple Rp = 0.1dB, the minimum stopband attenuation Rs = 40dB, and the stopband width BW = 4Hz.

8. The data fusion method for a distributed energy supply system according to claim 7, characterized in that: The transfer function of the prototype multi-notch filter is: Among them, H 1 (z), H 2 (z), H 3 (z), H 4 (z), H 5 (z) are the transfer functions of five prototype notch filters respectively, where z -1 , z -2 , …, z -30 is the unit impulse response sequence of the prototype notch filter, and a 1 , a 2 , …, a 30 and b 0 , b 1 , b 2 , …, b 30 are the coefficients corresponding to z 1 (z), H 2 (z), H 3 (z), H 4 (z), H 5 (z) after multiplication and calculation for the corresponding z -1 , z -2 , …, z -30 .

Images